time taken by memset

Hello,

I am trying to perform a simple experiment. I want to find out how much the memset will take in different embedded os's.

First of I tried to memset a malloced buffer of 2MB in wince 6.0 and I was able to do it in 5ms.

If I try the same in linux the memset takes around 10ms!! Almost double.

Now, I was wondering if caching would cause such huge numbers in linux. In that case how do I enable caching for my malloc'ed buffer in linux?

Any ideas or thoughts apperciated.